Organ segmentation device and method for magnetic resonance imaging, magnetic resonance imaging device

By combining localization image acquisition, coarse localization, and segmentation components, the problem of automatic planning in abdominal multi-organ scanning was solved, enabling rapid and accurate organ segmentation and localization, and improving the performance of magnetic resonance imaging devices.

CN115707429BActive Publication Date: 2026-11-10CANON MEDICAL SYST CORP
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202110959593.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-08-20
Publication Date
2026-11-10
Estimated Expiration
2041-08-20

AI Technical Summary

Technical Problem

Existing magnetic resonance imaging (MRI) techniques cannot achieve automatic planning in abdominal multi-organ scans, especially for asymmetric structures such as the liver, gallbladder, and pancreas. Furthermore, existing segmentation algorithms are time-consuming and cannot meet the requirements for fast and accurate scanning.

Method used

By combining localization image acquisition, coarse localization, and segmentation components, and through image segmentation algorithms and deep learning, abdominal organs are quickly segmented to provide their location and size information. Furthermore, by optimizing components for local optimization, precise localization and scan planning in three-dimensional space are achieved.

Benefits of technology

It enables rapid and precise segmentation and localization of multiple organs in the abdomen, shortens scanning time, improves the performance of magnetic resonance imaging devices, and supports efficient automatic scanning and 3D rendering.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115707429B_ABST
    Figure CN115707429B_ABST
Patent Text Reader

Abstract

This invention provides an organ segmentation device and method for magnetic resonance imaging (MRI), and an MRI device. The organ segmentation device for MRI includes: a positioning image acquisition unit that acquires three-dimensional or multi-layer two-dimensional positioning images of the organ; a coarse positioning unit that, based on the positioning images, coarsely positions the segment containing the organ in the layer direction; and a segmentation unit that performs image segmentation processing on each positioning image in the segment containing the organ to obtain the segmentation result of the organ. It enables rapid organ segmentation, fully utilizes rapid segmentation to provide complete information on the location and size of the organ; it is not limited by structural symmetry and can detect and segment asymmetrical structures; it can provide feature information such as the location and size of multiple organs in three-dimensional space and provide precise positioning; it standardizes and improves the accuracy of abdominal scanning, simplifies the operation process, reduces time, and improves the performance of the MRI device.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to an organ segmentation device and method for magnetic resonance imaging, an organ localization device for magnetic resonance imaging, an ROI (region of interest) planning device for organ scanning in magnetic resonance imaging, and an organ rendering device for magnetic resonance imaging. Background Technology

[0002] Magnetic resonance imaging (MRI) is an imaging technique commonly used in clinical practice. MRI uses high-frequency signals at the Larmor frequency to magnetically excite the atomic nuclei of a subject located in a static magnetic field, causing them to spin. The resulting nuclear magnetic resonance (NMR) signals are then used to reconstruct the image.

[0003] Clinically, abdominal MRI (hepatobiliary and pancreatic system) involves a series of complex multi-organ scans, requiring imaging information of digestive system organs such as the liver, gallbladder, and pancreas. Contrast enhancement may also be necessary in some cases. This series of scan planning tasks requires manual operation by technicians based on their individual skills and experience, which is time-consuming and cannot guarantee consistent and accurate scan placement; sometimes, rescanning is even required. Therefore, automated abdominal MRI scan planning has significant clinical value.

[0004] Existing automated planning techniques for magnetic resonance imaging scans, such as those utilizing structural symmetry, can be used for automated planning of the head, spine, knee joint, and shoulder joint. However, this method of utilizing structural symmetry cannot be applied to asymmetrical abdominal organs, and therefore cannot achieve automated planning of the abdomen.

[0005] In existing technologies, for example, methods for locating the liver include using the lower edge of the ribs as a substitute for the lower edge of the liver, but this method cannot accurately express the location of the lower edge of the liver.

[0006] Furthermore, in abdominal multi-organ scans, scan planning requires reference to the location and size information of multiple organs, such as the position of the diaphragm at the upper edge of the liver, and the positions of the pancreas and common bile duct in the case of the liver region. Additionally, to achieve efficient and even automated MRI scanning, abdominal scan planning must be precise, standardized, and rapid. Moreover, since the patient is inside the MRI scanner during the scan, the scan time needs to be as short as possible, thus placing high demands on scan planning time, requiring detection and automatic planning within seconds or even in real time. This, in turn, necessitates accelerating the segmentation and localization of the positions and sizes of various organs.

[0007] However, existing technologies applicable to the segmentation and localization of the location and size of abdominal organs are time-consuming when segmenting entire organs, such as the whole liver. Therefore, current segmentation algorithms for the location and size of organs cannot meet this requirement. Summary of the Invention

[0008] This invention addresses the aforementioned problems by providing a device and method for organ segmentation in magnetic resonance imaging (MRI), as well as an MRI device. This device enables rapid organ segmentation, fully utilizing the rapid segmentation to provide complete information on the location and size of organs, and achieving rapid localization. It is not limited by structural symmetry and can detect, segment, and locate asymmetrical structures. It provides feature information such as the location and size of multiple organs in three-dimensional space and provides precise localization. It enables automatic planning for scanning multiple organs in the abdomen, achieving standardization and precision in abdominal scanning, simplifying the operation process and reducing time. It achieves efficient automatic scanning planning, high-performance automatic scanning, and three-dimensional rendering, thereby improving the performance of the MRI device.

[0009] The organ segmentation device for magnetic resonance imaging provided by the present invention includes:

[0010] The localization image acquisition unit acquires three-dimensional or multi-layer two-dimensional localization images of organs;

[0011] The coarse positioning section, based on the positioning image, coarsely positions the segment where the organ is located in the layer direction; and

[0012] The segmentation unit performs image segmentation processing on each of the localized images in the segment where the organ is located, and obtains the segmentation result of the organ.

[0013] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0014] The coarse positioning unit, based on a search algorithm, selects two or more cross-sectional two-dimensional images from multiple cross-sectional two-dimensional images obtained from the positioning image, performs image segmentation processing on the selected cross-sectional two-dimensional images, determines the cross-sectional two-dimensional image corresponding to the end of the organ in the layer direction, and coarsely locates the segment where the organ is located in the layer direction.

[0015] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0016] The coarse positioning unit selects two or more cross-sectional two-dimensional images using an equal-interval selection method, a random selection method, or a selection method based on the distribution of the organs.

[0017] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0018] It also includes an optimization unit that optimizes local areas of the organ based on the organ segmentation results.

[0019] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0020] The optimization unit selects a local detection region based on the segmentation result of the organ, performs three-dimensional surface detection or two-dimensional edge detection based on the selected local detection region, calculates the local vertex position of the organ, and performs optimization.

[0021] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0022] The locality of the organ is any one of the six ends that define the location range of the organ.

[0023] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0024] The layer direction can be any of the following: head-to-toe, front-to-back, or left-to-right.

[0025] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0026] The segmentation results of the organs are data representing the outline, size, and location of the organs.

[0027] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0028] The search algorithm is any one of sequential search, binary search, tree structure search, or hash search.

[0029] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0030] The image segmentation process employs image segmentation algorithms or deep learning.

[0031] Furthermore, the organ segmentation device for magnetic resonance imaging provided by the present invention, wherein...

[0032] It also includes a body contour detection unit, which detects the body surface area in the cross-sectional two-dimensional images based on multiple cross-sectional two-dimensional images obtained from the positioning image.

[0033] The optimization unit optimizes the local area of ​​the organ based on the body surface area and the segmentation results of the organ.

[0034] In addition, the organ segmentation device for magnetic resonance imaging provided by this invention...

[0035] The organ in question is any one of the following organs: liver, kidney, pancreas, spleen, or heart.

[0036] Furthermore, the organ segmentation method in the organ segmentation device for magnetic resonance imaging provided by the present invention includes:

[0037] The step of acquiring localization images is to obtain three-dimensional or multi-layer two-dimensional localization images of the organs;

[0038] The coarse localization step, based on the localization image, coarsely localizes the segment where the organ is located in the layer direction;

[0039] The segmentation step involves performing image segmentation processing on each of the localized images in the segment where the organ is located to obtain the segmentation result of the organ.

[0040] Furthermore, the magnetic resonance imaging device provided by the present invention comprises:

[0041] The organ segmentation device for magnetic resonance imaging described in any of the above claims.

[0042] Furthermore, the magnetic resonance imaging device provided by the present invention comprises:

[0043] The organ segmentation device for magnetic resonance imaging described in any of the preceding claims; and

[0044] The positioning unit locates the organ based on the segmentation results of the organ.

[0045] Furthermore, the magnetic resonance imaging device provided by the present invention comprises:

[0046] The organ segmentation device for magnetic resonance imaging described in any of the preceding claims; and

[0047] The ROI planning department plans the location, orientation, and FOV size of the scanning ROI based on the segmentation results of the organ.

[0048] Furthermore, the magnetic resonance imaging device provided by the present invention comprises:

[0049] The organ segmentation device for magnetic resonance imaging described in any of the preceding claims; and

[0050] The rendering unit performs three-dimensional morphological rendering of the organ based on the segmentation results of the organ.

[0051] Technical effect

[0052] The organ segmentation device and method for magnetic resonance imaging (MRI) and the MRI device of the present invention can achieve rapid organ segmentation, fully utilize rapid segmentation to provide complete information on the location and size of organs, and achieve rapid localization; they can detect, segment, and locate asymmetrical structures without being limited by structural symmetry; they can provide feature information such as the location and size of multiple organs in three-dimensional space and provide precise localization; they can achieve automatic scanning planning for multiple organs in the abdomen, realize the standardization and precision of abdominal scanning, simplify the operation process and shorten the time, achieve efficient automatic scanning planning, high-performance automatic scanning and three-dimensional rendering, and improve the performance of the MRI device. Attached Figure Description

[0053] Figure 1 This is a schematic diagram illustrating an example of the structure of the organ segmentation device for magnetic resonance imaging according to the first embodiment of the present invention.

[0054] Figure 2 This is an operation flowchart of an organ segmentation device for magnetic resonance imaging according to the first embodiment of the present invention.

[0055] Figure 3 This is a schematic diagram of an example of a cross-sectional two-dimensional image selected by the coarse positioning section of the organ segmentation device for magnetic resonance imaging according to the first embodiment of the present invention.

[0056] Figure 4 This is a schematic diagram of an example of the segmentation unit of the organ segmentation device for magnetic resonance imaging according to the first embodiment of the present invention performing image segmentation processing on a cross-sectional two-dimensional image within a segment.

[0057] Figure 5 This is a schematic diagram of another example of the selection of a cross-sectional two-dimensional image by the coarse positioning section of the organ segmentation device for magnetic resonance imaging according to the first embodiment of the present invention.

[0058] Figure 6 This is a schematic diagram illustrating an example of the structure of the organ segmentation device for magnetic resonance imaging according to the second embodiment of the present invention.

[0059] Figure 7 This is an operation flowchart of an organ segmentation device for magnetic resonance imaging according to a second embodiment of the present invention.

[0060] Figure 8 This is a schematic diagram of a local feature of the optimization section of the organ segmentation device for magnetic resonance imaging according to the second embodiment of the present invention.

[0061] Figure 9 This is another schematic diagram showing the partial features of the optimization section of the organ segmentation device for magnetic resonance imaging according to the second embodiment of the present invention.

[0062] Figure 10 This is a schematic diagram of the optimization process performed by the optimization section of the organ segmentation device for magnetic resonance imaging according to the second embodiment of the present invention.

[0063] Figure 11 This is a schematic diagram illustrating an example of the structure of the organ segmentation device for magnetic resonance imaging according to the third embodiment of the present invention.

[0064] Figure 12 This is an operation flowchart of an organ segmentation device for magnetic resonance imaging according to the third embodiment of the present invention.

[0065] Figure 13 This is a schematic diagram of the body contour detection unit of the organ segmentation device for magnetic resonance imaging according to the third embodiment of the present invention performing detection processing.

[0066] Figure 14 This is a schematic diagram of an embodiment of the organ segmentation device for magnetic resonance imaging of the present invention.

[0067] Figure 15 This is yet another schematic diagram of an embodiment of the organ segmentation device for magnetic resonance imaging of the present invention.

[0068] Figure 16 This is a schematic diagram of a second embodiment of the organ segmentation device for magnetic resonance imaging of the present invention.

[0069] Figure 17 This is another schematic diagram of Embodiment 2 of the organ segmentation device for magnetic resonance imaging of the present invention.

[0070] Explanation of reference numerals in the attached figures

[0071] 100, 100a, 100b: Organ segmentation device for magnetic resonance imaging;

[0072] 101: Positioning image acquisition unit;

[0073] 102: Coarse positioning part;

[0074] 110: Segmentation section;

[0075] 103: Optimization Department;

[0076] 104: Body Contour Detection Department. Detailed Implementation

[0077] Hereinafter, with reference to the accompanying drawings, the structure and operation flow of the organ segmentation device for magnetic resonance imaging according to the embodiment will be described.

[0078] Furthermore, in this specification and accompanying drawings, only the components related to the technical concept of the present invention are described and shown, while other components are omitted.

[0079] In addition, in this specification and the accompanying drawings, components with the same or similar functions are labeled with the same reference numerals, and repeated descriptions are omitted where appropriate.

[0080] <First Implementation Method>

[0081] Structure of the organ segmentation device 100 for magnetic resonance imaging

[0082] Figure 1 This is a schematic diagram illustrating an example of the structure of the organ segmentation apparatus 100 for magnetic resonance imaging according to the first embodiment of the present invention.

[0083] Furthermore, in this invention, "organs" mainly refers to the organs in the abdomen, which may include the liver, kidneys, pancreas, and spleen, but are not limited to these; they may also include organs such as the heart. In this embodiment, the case of the liver as the organ will be used as an example for explanation.

[0084] like Figure 1 As shown, the organ segmentation device 100 for magnetic resonance imaging includes: a positioning image acquisition unit 101, a coarse positioning unit 102, and a segmentation unit 110.

[0085] The positioning image acquisition unit 101 acquires three-dimensional or multi-layer two-dimensional images of the organ as positioning images. For example, it can acquire a set of abdominal images in three-dimensional space as positioning images through scanning, where the FOV (Field of View) of the images includes the organ region (e.g., liver). Alternatively, the FOV can simply include the organ region or be larger than the organ region. Furthermore, this set of abdominal images in three-dimensional space can be multi-layer two-dimensional images or three-dimensional volumetric imaging images. Additionally, this set of abdominal images in three-dimensional space is a set of images arranged in a layer direction. This layer direction is the direction in which the scanned slices are arranged. Furthermore, this layer direction can be, for example, any direction of the human body, such as the head-to-toe direction, the front-back direction, or the left-right direction.

[0086] The coarse positioning unit 102, based on the positioning image, coarsely locates the segment where the organ is located in the layer direction. For example, the coarse positioning unit 102 may, based on a search algorithm, select two or more cross-sectional two-dimensional images from multiple cross-sectional two-dimensional images obtained from the positioning image, perform image segmentation processing on the selected cross-sectional two-dimensional images, determine the cross-sectional two-dimensional image corresponding to the end of the organ in the layer direction, and coarsely locate the segment where the organ is located in the layer direction.

[0087] Here, if the positioning image acquired by the positioning image acquisition unit 101 is a cross-sectional scan image, this positioning image can be used as a cross-sectional two-dimensional image. On the other hand, if the positioning image acquired by the positioning image acquisition unit 101 is not a cross-sectional scan image, multiple cross-sectional two-dimensional images are acquired by performing cross-sectional MPR (multi-plane reconstruction) based on the positioning image. The method for performing cross-sectional MPR (multi-plane reconstruction) based on the positioning image can be any known method and will not be elaborated here.

[0088] In addition, the coarse positioning unit 102 may, for example, select two or more cross-sectional two-dimensional images from multiple cross-sectional two-dimensional images obtained based on the positioning image using an equally spaced selection method, a random selection method, a selection method based on the distribution of organs, or other search algorithms.

[0089] In addition, regarding search algorithms, there are sorted and unsorted search algorithms such as sequential search, binary search, tree structure search, and hash search. Searches can also be performed based on the statistical patterns of liver location distribution.

[0090] The segmentation unit 110 performs image segmentation processing on each localization image within the segment containing the organ to obtain the organ segmentation result. Here, the image segmentation processing applies image segmentation algorithms or deep learning. Furthermore, the organ segmentation result is data representing the organ's contour, size, and location. For example, the segmentation unit 110 can perform image segmentation processing on each cross-sectional two-dimensional image along the layer direction from one end of the organ (e.g., the top layer of the liver) to the other end (e.g., the bottom layer of the liver), i.e., perform cross-sectional organ segmentation layer by layer to obtain the organ (e.g., the liver) segmentation result.

[0091] [Operation of the Organ Segmentation Device 100 for Magnetic Resonance Imaging]

[0092] Below, refer to Figures 2-5 The operation of the organ segmentation device 100 for magnetic resonance imaging is explained. Figure 2 This is an operation flowchart of the organ segmentation device 100 for magnetic resonance imaging according to the first embodiment of the present invention. Figure 3 This is a schematic diagram of an example of a cross-sectional two-dimensional image selected by the coarse positioning section 102 of the magnetic resonance imaging organ segmentation apparatus 100 according to the first embodiment of the present invention. Figure 4 This is a schematic diagram of an example of image segmentation processing performed by the segmentation unit 103 of the magnetic resonance imaging organ segmentation apparatus 100 according to the first embodiment of the present invention on a cross-sectional two-dimensional image within a segment. Figure 5 This is a schematic diagram of another example of the coarse positioning section 102 of the magnetic resonance imaging organ segmentation apparatus 100 according to the first embodiment of the present invention selecting a cross-sectional two-dimensional image.

[0093] like Figure 2 As shown, after the MRI organ segmentation device 100 starts operating, in step 100, the positioning image acquisition unit 102 acquires a three-dimensional or multi-layer two-dimensional abdominal positioning image of the organ. Here, for example, let's assume that the positioning image acquisition unit 102 acquires an image such as... Figure 3 Two-dimensional positioning images of the abdomen, showing 31 layers arranged in the layer direction, as shown on the left and right sides. Figure 3 Each layer in the representation is as follows: Figure 4 The image shown is divided into three sections on the left, top, middle, and bottom. Furthermore, since this is a two-dimensional cross-sectional image of the abdomen, the layer direction corresponds to the direction connecting the head and toes, i.e., the head-to-toe direction.

[0094] Next, in step 200, the coarse positioning unit 102 roughly locates the liver segment in the layer direction based on the abdominal positioning image. Specifically, for example, the coarse positioning unit 102 firstly selects a two-dimensional positioning image of the 31-layer cross-section obtained in step 100 as the probe layer; secondly, it performs liver image segmentation processing on the two-dimensional positioning image of the probe layer; and thirdly, based on the image segmentation processing results, it uses a search method to quickly locate the liver region, performs coarse positioning of the liver in the layer direction (here, the head-to-foot direction), defines the layer where the liver ends are located in the layer direction (i.e., the two-dimensional positioning image of the cross-section), and roughly locates the organ segment in the layer direction. The coarse positioning unit 102 can adjust the position of the probe layer and repeatedly execute steps one through three until the layer where the liver ends (liver top and liver base) are located in the layer direction is defined.

[0095] Here, the probe layer is a representative layer. The distribution of organs (here, the liver) can be determined based on the segmentation results of adjacent probe layers. For example, if the segmentation result of one probe layer in the layer direction does not include the organ region, while the segmentation result of the probe layer in the other layer direction includes the organ region, it can be determined that one end of the organ in the layer direction is located in a layer between these two probe layers. In this case, the segment defined by these two adjacent probe layers is considered a present segment. If the segmentation result of one probe layer in the layer direction includes the organ region, while the segmentation result of the probe layer in the other layer direction does not include the organ region, it can be determined that the other end of the organ in the layer direction is located in a layer between these two probe layers. In this case, the segment defined by these two adjacent probe layers is considered a absent segment. Furthermore, if the segmentation results of two adjacent probe layers in the layer direction both do not include the organ region, it can be determined that the organ does not exist in any layer between these two probe layers. In this case, the segment defined by these two adjacent probe layers is considered a non-existent segment. In addition, if the segmentation results of two adjacent probe layers in the layer direction both include the organ region, it can be determined that the organ exists in any layer between these two probe layers. In this case, the segment defined by the two adjacent probe layers is regarded as the existing segment.

[0096] Regarding the selection of the probe layer, it is preferably selected from the appearance and disappearance segments. Further searching within these segments using a search algorithm can determine the top and bottom layers of the liver. Furthermore, various selection methods are available; for example, selection can be made randomly from multiple cross-sectional two-dimensional images obtained from the localization image, at equal intervals, based on organ distribution, or based on other data search algorithms. The search algorithm can be any of the following: sequential search, binary search, tree structure search, or hash search.

[0097] Next, in step 300, based on the segment where the organ is located in the layer direction, i.e. the segment from the top layer of the liver to the bottom layer of the liver, which was roughly located in step 200, the liver is segmented layer by layer (two-dimensional positioning image of cross section) for this segment to obtain the segmentation result.

[0098] The following is combined Figure 3 and Figure 4 An application example of the operation of the magnetic resonance imaging organ segmentation apparatus 100 according to the first embodiment will be described. In this application example, for example, the positioning image acquisition unit 102 acquires a two-dimensional positioning image of the abdomen in cross-section of 31 layers arranged in the layer direction, where the layer direction corresponds to the head-to-foot direction, with the aim of segmenting the liver. Furthermore, Figure 3 Each layer in the representation is as follows: Figure 4An image like the one shown in the top, middle, and bottom sections of the left column. Figure 3 and Figure 4 In the diagram, dashed layers represent layers selected as probe layers, and dotted-dash layers represent layers whose image segmentation results include organs.

[0099] In this application example, such as Figure 3 As shown, the positioning image acquisition unit 102 acquired a two-dimensional positioning image of the abdomen of a cross-section of 31 layers arranged in the layer direction, and selected 6 probe layers in an equally spaced manner. Figure 3 The layer shown by the dashed line on the left in the middle is used to segment the probe layer, that is, to segment it as shown in the image. Figure 4 The two-dimensional cross-sectional image shown on the left is segmented, and the probe layers that segment the liver region are marked, here as dashed-dot layers. The result of this segmentation process is that the segmentation results of 3 out of 6 probe layers include the liver region. Figure 3 (The dotted-dash layer on the right side of the middle). That is, as shown... Figure 3 As shown on the right, from top to bottom along the layer direction, the segments are: appearance segment, existence segment, existence segment, and disappearance segment. Next, probe layers are further set up for the appearance and disappearance segments, and a search algorithm is used to determine the top layer of the liver (…). Figure 4 The first dotted-line layer from the middle to the upper part) and the liver basal layer ( Figure 4 The first layer is marked with a dotted line (from the bottom). Next, each layer within the segment defined by the top and bottom layers of the liver is segmented to obtain information such as the liver's outline, location, and size in three-dimensional space (refer to...). Figure 4 The middle images are used as the segmentation result.

[0100] Thus, according to the organ segmentation apparatus 100 for magnetic resonance imaging according to the first embodiment, the positioning image acquisition unit 102 acquires three-dimensional or multi-layer two-dimensional positioning images of the organ. Based on the positioning images, the coarse positioning unit 102 coarsely positions the segment where the organ is located in the layer direction. Then, the segmentation unit 103 performs image segmentation processing on each positioning image in the segment where the organ is located to obtain the segmentation result of the organ. Thus, through the processing of the coarse positioning unit 102, the segment where the organ is located in the layer direction is coarsely positioned with a small amount of image segmentation processing. Then, image segmentation processing is performed on each cross-sectional two-dimensional image within the segment where the organ is located. Therefore, the organ can be quickly coarsely positioned with a small amount of computation, the range of image segmentation processing required can be reduced, the amount of computation required to segment the organ can be reduced, the processing speed can be increased, and the processing time can be shortened. On the other hand, by performing image segmentation processing on each cross-sectional two-dimensional image within the segment where the organ is located to obtain the organ segmentation result, more complete spatial information such as the position and size of the organ can be provided. That is, the organ segmentation device 100 for magnetic resonance imaging according to the first embodiment can quickly provide more complete spatial information such as the location and size of organs; can achieve rapid positioning; can detect, segment and locate asymmetrical structures without being limited by structural symmetry; can provide feature information such as the location and size of multiple organs in three-dimensional space and accurate positioning; and can improve the performance of magnetic resonance imaging devices.

[0101] in addition, Figure 3 The image shows an example of selecting probe layers at equal intervals to determine the liver segment based on image segmentation results and a search algorithm, but it is not limited to this; other methods can also be used... Figure 5 As shown, instead of selecting probe layers at equal intervals, the distribution of organs along the layer direction is considered. More probe layers are selected for sections where the organs are more likely to be present at their ends, and fewer probe layers are selected for sections where the organs are more likely to appear consecutively or where the organs are less likely to be present. This allows for rapid coarse localization of organs with less computation, further narrowing the area requiring image segmentation, reducing the computational load for organ segmentation, increasing processing speed, and shortening processing time.

[0102] <Second Implementation Method>

[0103] Structure of the organ segmentation device 100a for magnetic resonance imaging

[0104] Figure 6This is a schematic diagram illustrating an example of the structure of the organ segmentation apparatus 100a for magnetic resonance imaging according to the second embodiment of the present invention. In this embodiment, the case of the liver as the organ will be described as an example, similar to the first embodiment. In addition, in this embodiment, the same reference numerals are used for components that are the same or similar to those in the first embodiment, and detailed descriptions of these components are omitted; only the differences are described in detail.

[0105] like Figure 6 As shown, the organ segmentation device 100 for magnetic resonance imaging includes: a positioning image acquisition unit 101, a coarse positioning unit 102, an optimization unit 103, and a segmentation unit 110.

[0106] The positioning image acquisition unit 101, coarse positioning unit 102, and segmentation unit 110 are the same as in the first embodiment, and will not be described in detail here.

[0107] The optimization unit 103 optimizes the local area of ​​the organ based on the organ segmentation results. Here, the local area of ​​the organ refers to any one of the six ends that define the location range of the organ. Alternatively, the optimization unit may, for example, select a local detection region based on the organ segmentation results, perform three-dimensional surface detection or two-dimensional edge detection based on the selected local detection region, calculate the vertex position of the local area of ​​the organ, and then perform optimization. Here, the vertex can be, for example, a boundary point obtained by using the center, centroid, or other weight calculation methods of the local detection region.

[0108] Furthermore, taking the liver as an example, because the liver is close to the diaphragm, it is affected by the movement of the diaphragm. Therefore, when considering the influence of respiration on the position of the liver apex, the requirements for the edge / boundary points are relatively high, requiring more refined detection of the liver apex region to accurately determine the position of the apex. Additionally, when determining the morphological size of the liver, such as determining hepatomegaly requiring measuring the distance between the liver apex and base, more refined detection of the liver base is also needed to accurately determine the position of the apex of the liver base region. The optimization unit 103 is preferably applied in the above-described cases.

[0109] [Operation of the Magnetic Resonance Imaging Organ Segmentation Device 100a]

[0110] Below, refer to Figures 7-10 The operation of the organ segmentation device 100a for magnetic resonance imaging is explained. Figure 7 This is an operation flowchart of the organ segmentation device 100a for magnetic resonance imaging according to the second embodiment of the present invention. Figure 8 This is a schematic diagram of a local feature of the optimization unit 103 of the magnetic resonance imaging organ segmentation apparatus 100a according to the second embodiment of the present invention, which is subjected to optimization processing. Figure 9This is another schematic diagram of the local features of the optimization unit 103 of the magnetic resonance imaging organ segmentation device 100a according to the second embodiment of the present invention, which is subjected to optimization processing. Figure 10 This is a schematic diagram of the optimization unit 103 of the magnetic resonance imaging organ segmentation device 100a according to the second embodiment of the present invention performing optimization processing.

[0111] Figure 7 Steps 100-300 in the middle Figure 2 Steps 100 to 300 are the same, so detailed descriptions are omitted here.

[0112] like Figure 7 As shown, in step 300, the organ segmentation device 100 for magnetic resonance imaging performs liver segmentation layer by layer according to the segment where the organ is located in the layer direction, i.e., the segment from the top layer of the liver to the bottom layer of the liver, which is roughly located in step 200, and obtains the segmentation result. Then, without ending the process, it proceeds to step S400.

[0113] In step S400, the optimization unit 103 optimizes the local area of ​​the liver based on the liver segmentation results in step S300.

[0114] Here, a localized area of ​​the liver is, for example... Figure 8 and Figure 9 The diagram shows any one of the six ends that define the location of the liver: the anterior end (A), the posterior end (P), the right end (R), the left end (L), the apex (also known as the liver top boundary point) (H), and the basal end (also known as the liver bottom boundary point) (F).

[0115] Specifically, the optimization unit 103 can first extract the liver apex boundary point H based on the liver segmentation results; then, using point H as a reference point, and selecting a fine detection area of ​​the liver apex as a local detection area according to the size and location of the patient's liver region and the size and location of the body contour; next, within the selected local detection area, performing three-dimensional surface detection or two-dimensional edge detection to reconstruct a surface as the top surface of the liver; then, based on a threshold set according to clinical accuracy requirements, selecting the top surface of the liver or a set of liver apex surfaces composed of several adjacent layers; finally, based on the selected liver apex surface or set of liver apex surfaces, comprehensively calculating the center position of the liver apex to complete the optimization. Here, the specific calculation method for calculating the center position of the liver apex can be a method such as calculating the center, centroid, or weighted average of multiple points.

[0116] Here, the example of the liver apex boundary point H is used, but the optimization unit 103 can apply the same method to optimize other ends of the liver.

[0117] Thus, the organ segmentation apparatus 100a for magnetic resonance imaging according to the second embodiment, in addition to having the effects of the first embodiment, has an optimization unit that performs more detailed detection on the local parts of the organ based on the organ segmentation results, thereby providing more accurate and detailed feature information such as the position and size of the organ in three-dimensional space and precise positioning, further improving the performance of the magnetic resonance imaging apparatus.

[0118] <Third Implementation Method>

[0119] Structure of the organ segmentation device 100b for magnetic resonance imaging

[0120] Figure 11 This is a schematic diagram illustrating an example of the structure of the organ segmentation apparatus 100b for magnetic resonance imaging according to the third embodiment of the present invention. In this embodiment, the case of the liver as the organ is described as an example, similar to the first and second embodiments. In addition, in this embodiment, the same reference numerals are used for components that are the same or similar to those in the first embodiment, and detailed descriptions of these components are omitted; only the differences are described in detail.

[0121] like Figure 11 As shown, the organ segmentation device 100 for magnetic resonance imaging includes: a positioning image acquisition unit 101, a coarse positioning unit 102, an optimization unit 103, a body contour detection unit 104, and a segmentation unit 110.

[0122] The positioning image acquisition unit 101, coarse positioning unit 102, and segmentation unit 110 are the same as in the second embodiment, and will not be described in detail here.

[0123] The body contour detection unit 104 detects the body surface area in the cross-sectional two-dimensional images based on multiple cross-sectional two-dimensional images acquired from the positioning image.

[0124] In this embodiment, the optimization unit 103 optimizes the local area of ​​the organ based on the body surface area and the segmentation result of the organ.

[0125] Operation of the organ segmentation device 100b for magnetic resonance imaging

[0126] Below, refer to Figures 12-13 The operation of the organ segmentation device 100b for magnetic resonance imaging is explained. Figure 12 This is an operation flowchart of the organ segmentation device 100b for magnetic resonance imaging according to the third embodiment of the present invention. Figure 13 This is a schematic diagram of the body contour detection unit 104 performing detection processing in the organ segmentation apparatus 100b for magnetic resonance imaging according to the third embodiment of the present invention.

[0127] Figure 12Steps 100-300 in the middle Figure 2 Steps 100 to 300 are the same, so detailed descriptions are omitted here.

[0128] like Figure 12 As shown, in step S500, the body contour detection unit 104 detects the body surface area in the cross-sectional two-dimensional images based on multiple cross-sectional two-dimensional images acquired according to the positioning image. For example, as Figure 13 As shown, the green box is detected as the body surface area in the cross-sectional two-dimensional image. Here, various methods, such as well-known edge detection methods or grayscale analysis methods, can be used to detect the body surface area in the cross-sectional two-dimensional image, which will not be elaborated here.

[0129] In step 400, the optimization unit 103 optimizes the local area of ​​the liver based on the liver segmentation results in step S300 and the body surface area in the cross-sectional two-dimensional image in step 500.

[0130] Thus, the organ segmentation apparatus 100b for magnetic resonance imaging according to the third embodiment, in addition to having the effects of the first embodiment, has a body contour detection unit 104 that detects the body contour on the cross-section, can remove the arm area, and can quickly and efficiently locate the body's front-back-left-right (APLR) center and the range of effective anatomical structures. This removes data that interferes with the optimization unit 103, and can further provide more accurate and detailed feature information such as the position and size of organs in three-dimensional space and precise positioning, further improving the performance of the magnetic resonance imaging apparatus.

[0131] The above description, using the liver as an example, illustrates several embodiments of the present invention, but it is not limited thereto; the technical concept of the present invention can also be implemented in other ways. The following description, in conjunction with Embodiment 1 and Embodiment 2, will further illustrate these embodiments.

[0132] Furthermore, Embodiments 1 and 2 can employ any of the structures described in the first to third embodiments, along with corresponding flowcharts. The following description uses the structure described in the first embodiment and the corresponding flowchart.

[0133] (Example 1)

[0134] Reference Figure 14 and Figure 15 Example 1 will be described below. Figure 14 This is a schematic diagram of an embodiment of the organ segmentation device for magnetic resonance imaging of the present invention. Figure 15 This is yet another schematic diagram of an embodiment of the organ segmentation device for magnetic resonance imaging of the present invention.

[0135] In this first embodiment, as Figure 14As shown, the positioning image acquisition unit 102 acquires three-dimensional or multi-layer two-dimensional abdominal positioning images of the organs. Here, for example, let's assume that the positioning image acquisition unit 102 acquires images such as... Figure 14 Two-dimensional positioning images of the abdomen, showing multiple cross-sections arranged in the layer direction as shown on the left and right sides. Figure 14 Each layer in the representation is as follows: Figure 15 An image as shown on the left. Furthermore, since this is a two-dimensional cross-sectional image of the abdomen, the layer direction corresponds to the direction connecting the head and toes, i.e., the head-to-toe direction. In Example 1, the segmentation of the kidney will be used as an example for explanation.

[0136] In detail, in Embodiment 1, after the positioning image acquisition unit 102 acquires a two-dimensional positioning image of the abdomen in multiple cross-sectional layers arranged in the layer direction, since the kidneys are mostly distributed in the lower part of the abdominal data (e.g., the lower 1 / 2 to 1 / 3), a denser probe layer is selected for the lower 1 / 2 to 1 / 3 of the positioning image of the abdomen in multiple cross-sectional layers arranged in the layer direction, and a sparser probe layer is selected for other locations. Figure 14 The layer shown by the dashed line on the left in the middle is used to segment the probe layer, that is, to segment it as shown in the image. Figure 15 The two-dimensional cross-sectional image shown on the left is segmented, and the probe layers that segment the kidney region are marked, here as dashed-dot layers. The result of this segmentation process is that the segmentation results of 3 out of 6 probe layers include the kidney region. Figure 14 (The dotted-dash layer on the right side of the middle). That is, as shown... Figure 14 As shown on the right, from top to bottom along the layer direction, the segments are: non-existent segment, non-existent segment, present segment, present segment, present segment, disappearing segment, and non-existent segment. Next, probe layers are further set up for the present and disappearing segments, and a search algorithm is used to determine the top layer of the kidney (…). Figure 14 The first dotted-line layer from the middle to the upper part) and the basal layer of the kidney ( Figure 14 The first dotted-line layer starts from the middle and lower part of the kidney. Next, each layer within the segment defined by the apex and basal layer of the kidney is subdivided (see [reference]). Figure 15 (in the dashed-dot layer), to obtain information such as the outline, location, and size of the kidney in three-dimensional space (refer to...) Figure 15 The middle image is used as the segmentation result.

[0137] Thus, according to this first embodiment, it has the effects of the first embodiment, and can more accurately select the probe layer based on the statistical law of the kidney location distribution. In this way, it can quickly and coarsely locate the organ with less computation, further reduce the range that needs to be image segmented, further reduce the amount of computation required to segment the organ, further improve the processing speed, and further shorten the processing time.

[0138] (Example 2)

[0139] Reference Figure 16 and Figure 17 Example 2 will be described below. Figure 16 This is a schematic diagram of a second embodiment of the organ segmentation device for magnetic resonance imaging of the present invention. Figure 17 This is another schematic diagram of Embodiment 2 of the organ segmentation device for magnetic resonance imaging of the present invention.

[0140] In this second embodiment, as Figure 15 As shown, the positioning image acquisition unit 102 acquires three-dimensional or multi-layer two-dimensional abdominal positioning images of the organs. Here, for example, let's assume that the positioning image acquisition unit 102 acquires images such as... Figure 16 The left and right sides show two-dimensional localization images of the abdomen in the coronal plane arranged in the layer direction, as shown. Figure 16 Each layer in the representation is as follows: Figure 17 An image as shown on the left. Furthermore, since this is a two-dimensional localization image of the abdomen in the coronal plane, the layer direction corresponds to the direction connecting the front and back of the body, i.e., the anterior-posterior direction. In Example 2, the segmentation of the kidney will be used as an example for explanation.

[0141] In detail, in Embodiment 2, after the positioning image acquisition unit 102 acquires a two-dimensional positioning image of the abdomen in a coronal plane with multiple layers arranged in the layer direction, since the kidneys are mostly distributed in the lower part of the abdominal data (e.g., the lower 2 / 5), a denser probe layer is selected for the 2 / 5 of the two-dimensional positioning image of the abdomen in a coronal plane arranged in the layer direction from back to front, and a sparser probe layer is selected for other locations. Figure 16 The layer shown by the dashed line on the left in the middle is used to segment the probe layer, that is, to segment it as shown in the image. Figure 17 The coronal 2D image shown on the left is segmented, and the probe layers that segment the kidney region are marked, here as dashed-dot layers. The result of this segmentation process is that two out of the five probe layers segment the kidney region. Figure 16 (The dotted-dash layer on the right side of the middle). That is, as shown... Figure 16 As shown on the right, from front to back along the layer direction, the segments are: non-existent segment, present segment, present segment, disappearing segment, and non-existent segment. Next, probe layers are further set up for the present and disappearing segments, and a search algorithm is used to determine the anterior layer of the kidney (…). Figure 17 The first dotted-line layer from the middle left) and the posterior layer of the kidney ( Figure 17 The first dotted-line layer from the right in the middle). Next, each layer within the segment defined by the anterior and posterior layers of the kidney is subdivided (see...). Figure 17(in the dashed-dot layer), to obtain information such as the outline, location, and size of the kidney in three-dimensional space (refer to...) Figure 17 The middle image is used as the segmentation result.

[0142] Thus, according to this second embodiment, it has the effects of the first embodiment and the first embodiment described above, and it can quickly segment and locate the kidney by utilizing the coronal plane positioning.

[0143] As described above, an embodiment of the present invention implemented using an organ segmentation apparatus and method for magnetic resonance imaging has been presented. Furthermore, the present invention can also be implemented using a magnetic resonance imaging apparatus.

[0144] For example, the present invention can provide a magnetic resonance imaging apparatus comprising: the organ segmentation device for magnetic resonance imaging as described in any of the preceding claims.

[0145] Furthermore, the present invention can provide a magnetic resonance imaging (MRI) device comprising: the organ segmentation device for MRI as described in any one of the preceding claims; and a positioning unit for positioning the organ based on the segmentation result. According to the MRI device, rapid positioning can be achieved; it can detect, segment, and position asymmetrical structures without being limited by structural symmetry; it can provide feature information such as the position and size of multiple organs in three-dimensional space and provide precise positioning, thereby improving the performance of the MRI device.

[0146] Furthermore, the present invention can provide a magnetic resonance imaging (MRI) device comprising: the organ segmentation device for MRI as described in any of the preceding claims; and a scanning ROI planning unit that plans the position, orientation, and FOV size of the scanning ROI based on the organ segmentation results. According to the MRI device, automatic scanning planning of multiple abdominal organs can be achieved, realizing the standardization and precision of abdominal scanning, simplifying the operation process and reducing time, enabling efficient automatic scanning planning and high-performance automatic scanning, and improving the performance of the MRI device.

[0147] Furthermore, the present invention can provide a magnetic resonance imaging apparatus comprising: an organ segmentation device for magnetic resonance imaging as described in any of the preceding claims; and a rendering unit that performs three-dimensional morphological rendering of the organ based on the segmentation result. According to the magnetic resonance imaging apparatus, three-dimensional rendering can be achieved, thereby improving the performance of the magnetic resonance imaging apparatus.

[0148] Furthermore, the magnetic resonance device embodying the technical concept of the present invention, having an organ segmentation device for magnetic resonance imaging, also possesses the technical effects of the aforementioned organ segmentation device for magnetic resonance imaging.

[0149] The foregoing has described several embodiments of the present invention, but these embodiments are merely illustrative and are not intended to limit the scope of the invention. These new embodiments can be implemented in various other ways, and various omissions, substitutions, and modifications can be made without departing from the scope of the invention's spirit. These embodiments and their variations are included in the scope and spirit of the invention, and are also included in the scope of the invention and its equivalents described in the technical solution.

Claims

1. An organ segmentation device for magnetic resonance imaging, comprising: The localization image acquisition unit acquires three-dimensional or multi-layer two-dimensional localization images of organs; The coarse positioning unit selects multiple cross-sectional two-dimensional images from multiple cross-sectional two-dimensional images acquired based on the positioning image, determines a first cross-sectional two-dimensional image containing one end of the organ in the layer direction and a second cross-sectional two-dimensional image containing the other end of the organ in the layer direction from the selected multiple cross-sectional two-dimensional images, and determines the segment defined by the first cross-sectional two-dimensional image and the second cross-sectional two-dimensional image, thereby coarsely positioning the segment where the organ is located in the layer direction; as well as The segmentation unit performs image segmentation processing on each of the localized images in the segment where the organ is located, and obtains the segmentation result of the organ.

2. The organ segmentation device for magnetic resonance imaging as described in claim 1, wherein, The coarse positioning unit, based on a search algorithm, selects two or more cross-sectional two-dimensional images from multiple cross-sectional two-dimensional images obtained from the positioning image, performs image segmentation processing on the selected cross-sectional two-dimensional images, determines the first cross-sectional two-dimensional image and the second cross-sectional two-dimensional image, and coarsely locates the segment where the organ is located in the layer direction.

3. The organ segmentation device for magnetic resonance imaging as described in claim 2, wherein, The coarse positioning unit selects two or more cross-sectional two-dimensional images using an equal-interval selection method, a random selection method, or a selection method based on the distribution of the organs.

4. The organ segmentation device for magnetic resonance imaging as described in claim 1, wherein, It also includes an optimization unit that optimizes local areas of the organ based on the organ segmentation results.

5. The organ segmentation device for magnetic resonance imaging as described in claim 4, wherein, The optimization unit selects a local detection region based on the segmentation result of the organ, performs three-dimensional surface detection or two-dimensional edge detection based on the selected local detection region, calculates the local vertex position of the organ, and performs optimization.

6. The organ segmentation device for magnetic resonance imaging as described in claim 4, wherein, The locality of the organ is any one of the six ends that define the location range of the organ.

7. The organ segmentation device for magnetic resonance imaging as described in claim 1, wherein, The layer direction can be any of the following: head-to-toe, front-to-back, or left-to-right.

8. The organ segmentation device for magnetic resonance imaging as described in claim 1, wherein, The segmentation results of the organs are data representing the outline, size, and location of the organs.

9. The organ segmentation device for magnetic resonance imaging as described in claim 3, wherein, The search algorithm is any one of sequential search, binary search, tree structure search, or hash search.

10. The organ segmentation device for magnetic resonance imaging as described in claim 1, wherein, The image segmentation process employs image segmentation algorithms or deep learning.

11. The organ segmentation device for magnetic resonance imaging as described in claim 4, wherein, It also includes a body contour detection unit, which detects the body surface area in multiple cross-sectional two-dimensional images acquired based on the positioning image. The optimization unit optimizes the local area of ​​the organ based on the body surface area and the segmentation results of the organ.

12. The organ segmentation apparatus for magnetic resonance imaging as described in any one of claims 1 to 11, The organ in question is any one of the following organs: liver, kidney, pancreas, spleen, or heart.

13. A method for organ segmentation in an organ segmentation device for magnetic resonance imaging, comprising: The step of acquiring localization images is to obtain three-dimensional or multi-layer two-dimensional localization images of the organs; The coarse positioning step involves selecting multiple cross-sectional two-dimensional images from the multiple cross-sectional two-dimensional images obtained based on the positioning image, determining a first cross-sectional two-dimensional image containing one end of the organ in the layer direction and a second cross-sectional two-dimensional image containing the other end of the organ in the layer direction from the selected multiple cross-sectional two-dimensional images, and determining the segment defined by the first cross-sectional two-dimensional image and the second cross-sectional two-dimensional image, thereby coarsely positioning the segment where the organ is located in the layer direction; The segmentation step involves performing image segmentation processing on each of the localized images in the segment where the organ is located to obtain the segmentation result of the organ.

14. A magnetic resonance imaging device, comprising: The organ segmentation apparatus for magnetic resonance imaging according to any one of claims 1 to 12.

15. A magnetic resonance imaging device, comprising: The organ segmentation apparatus for magnetic resonance imaging according to any one of claims 1 to 12; and The positioning unit locates the organ based on the segmentation results of the organ.

16. A magnetic resonance imaging device, comprising: The organ segmentation apparatus for magnetic resonance imaging according to any one of claims 1 to 12; and The ROI planning department plans the location, orientation, and FOV size of the scanning ROI based on the segmentation results of the organ.

17. A magnetic resonance imaging device, comprising: The organ segmentation apparatus for magnetic resonance imaging according to any one of claims 1 to 12; and The rendering unit performs three-dimensional morphological rendering of the organ based on the segmentation results of the organ.

Citation Information

Patent Citations

  • Magnetic resonance imaging apparatus

    US20180180695A1

  • Magnetic resonance imaging system and method

    US20210158563A1